The Cost of Change
Reviewed by Greg Wilson / 2023-05-18
Keywords: Editorial
Everybody wants things to get better. People who are serious about it draw up budgets, because that forces them to be explicit about exactly what they're going to do and what's needed to make it happen.
I'm sure it's clear by now that we'd like more programmers to know what software engineering researchers have learned over the past fifty years and how to act on it. We have a plan to make that happen: replace the team programming project course that most students do in the third or fourth year of an undergraduate degree in computer science with one in which they work in teams to design an experiment, collect data, analyze it, and figure out what (if anything) they've proven.
- It's a minimally disruptive adjustment to the curriculum, so other faculty won't object on administrative grounds.
- it gives CS students a chance to learn some practical data science (which most of them will enjoy and which will make them more employable).
- It helps prepare them for graduate school (so it's in professors' own interests as well).
- And crucially, they will be more likely to understand and value researchers' experimental findings if they have done an experiment of their own.
What we need to make this happen is two dozen lectures' worth of material as both prose and slides with examples, exercises, and grading rubrics so that busy professors don't each have to build it all from scratch. Figuring a week of research and development per lecture (because most of this material would be genuinely new), then doubling that estimate because I'm always optimistic, it would take one person-year to build this course, run it for the first time, fix what didn't work, and package it up for general use.
If two dozen professors in software engineering want this to happen and will each contribute US$6500 to fund two weeks' worth of work, our entire discipline will have what it needs to shape young minds today so that tomorrow's programmers will be more engaged with our work. If you're at ICSE'23 in Melbourne right now, please chat with your colleagues about this idea. If your answer is "no", please mail me and tell me what your alternative plan is. "If we just keep publishing, surely they'll start to pay attention" has been failing for fifty years; I think it's time we tried something else.